Faster Onboarding
Thanks to the simpler integration of SSO and CLI, the onboarding time was reduced by 60%.
Client Overview
Client: AI Model Registry & Edge Inference Platform
Location: United States
Industry: Artificial Intelligence and Edge Computing, Enterprise SaaS
Project Background
AI Model Hub is a secure cloud-based platform that allows the management, storage, and utilization of AI models, especially in the case of commercial and industrial applications. We incorporated the capability of safe model publishing, sharing, and usage. The platform features enterprise-level licensing management that supports edge deployment, offline access, and secure role-based access. The four primary elements are, Studio: A web-based application where people can test and fine tune, Hub: Central Repository where models are stored, Server: A command-line interface (CLI) for model inference and optimization, and Auth: Central auth between above all three things.
Technical Challenges
Mixed-Platform RBAC: Developing an access system to operate across web, CLI, and API with Auth0 integration for authentication and SSO behind all three.
Retrievable Data Transfers: Allowing uploading and downloading of several gigabytes at once through GCS and S3, along with resumable, fault-tolerant transfer logic.
Versioning & Disposal: Developing an automatic model versioning system with configurable retention and disposal policies for storage costs optimization.
CLI and Web Parity: Achieving no difference between the web and command-line interfaces for publishing, fine-tuning, and model retrieving processes.
Technical Implementation
Auth0 Centralized Authentication
A blend of unified SSO, token-based access, and API-level authentication. For enterprises, RBAC ensures controlled model visibility while detailed audit logs maintain transparency and compliance across all environments.
GCP Serverless
Architecture
Cloud Functions were used to manage fine-tuning pipelines, model versioning, and artifact uploads. The serverless setup automatically scales to handle workload demands while minimizing infrastructure overhead.
Resumable Storage with GCS/S3
Implemented resumable uploads and downloads with intelligent caching, enabling reliable transfer of multi-gigabyte models. The setup ensures seamless access even in low-bandwidth or offline environments.
Unified CLI + Web Ecosystem
Built a unified push, search, and pull interface ensuring identical functionality across CLI and web. Developers can manage, deploy, and retrieve AI models seamlessly through a consistent command set.
Business Benefits
Enhanced Collaboration
Efficient and secure inter-departmental collaboration was ensured by implementing role-based access control (RBAC).
Scalable Storage
Version labeling was applied to the management of petabyte-scale model storage, enabling efficient retrieval and updates.
Accelerated Model Iteration
Automated dataset pipelines reduced model iteration cycles from weeks to days, improving development speed and deployment agility.
Edge Compliance
Enforced license and policy compliance on edge devices to ensure secure and compliant AI model deployment across enterprise systems.
Key Innovation
The combination of modular architecture lets you start small and then scale up. Developers are attracted to the project because of the uniform CLI and UI experiences. Controlling cloud expenses through deletion and retention should be an early consideration. Public-private model governance is a way to balance marketing and security of IP.
Solution
The partnership between Coditude and the client produced a next-generation AI Model Hub, a platform that is secure, scalable, and edge-ready that enables the enterprises to easily manage, fine-tune, deploy, and version AI models.
Coditude’s implementation of fine-tuning pipelines, enterprise-grade RBAC, and flexible offline capabilities provided a solid foundation for the client’s global AI infrastructure, a practical alternative to the public model registries like Hugging Face, with the necessary governance.
They loved working with Coditude!
We act as a global catalyst for change, empowering our clients and partners to leverage technology, unlock growth, achieve sustainability, and bolster capabilities. See how our clients benefited from our partnerships.
Impact Stories: From Vision to Product
Case Study
Pioneering Progress
The client, a market leader in immigration and visa services with a vast network spanning 60 offices in 27 countries, sought to revamp their outdated digital infrastructure.
View Case Study
